Phonology Orthography

Old Scandinavian is recorded in two different scripts, the runic script (the Futhark) and the Roman alphabet, which came into use with the introduction of Christianity shortly after the turn of the millennium. With certain additions the latter was made quite suitable as a means of representing the sounds and phonemes of Old Norse. The <1» and later the <5> for the voiced counterpart were borrowed from Old English. The <y> for the front, high labial vowel was also borrowed from Anglo-Saxon. Digraphs were used to represent the

Table 3.1 Vowel phonemes of Old Norse i i yyuüe^0oeoöaeaä9 High + + + + + + -- -- -- -- -Low - -- -- -- -- -- - + + + + Back ---- + + -- -- + + - + + + Labial - - + + + + -- + + + + -- - + Long - + - + - + - + - + - + + - + -

rich vowel system of Old Scandinavian. In addition various diacritics were occasionally adopted both for vowel quality and quantity. In the standardized spelling used in edited texts and adopted here, the acute accent' is used to denote long vowels. The letters used and their phonetic value can be seen from Tables 3.1 and 3.3.

Vowels

Old Norse Vowel System

The vowel phonemes of Old Norse can be represented as in Table 3.1, where the vowels are given in the standard orthography. The main redundancy in the system is that non-low back vowels are always labial. There are seven pairs distinguished by length only. Early in the period the short /ae/ merged with /e/. The long variant of /<?/ merged with lil early in the thirteenth century, and is represented by that letter in most of the classical texts. In a later development, in Norwegian and Swedish, the labial l&l also tended to become higher, and thus it would come closer to /6/. This vowel would in turn move up and threaten to merge with /u/, which then would move forward and become a high central vowel.

The N and the /u/ can also occur in a non-syllabic position and function as semivowels, /j/ and /w/ (the latter written <v>). In Ancient Scandinavian, /j/ was lost word initially, dr 'year' (< *jara)9 and /w/ was lost in front of stressed labial vowels, ulfr 'wolf (< *wulfaz).

This vowel system has evolved from the Ancient Scandinavian system through the process of umlaut. Ancient Scandinavian had the five canonical vowels /i, u, e, o, a/, which could be long or short. In stressed syllables preceding unstressed syllables with the vowel N (syllabic or semivowel) the back vowels would have a fronted allophone: /u/ > [y], /o/ > [0], /a/ > [ae], /au/ > [ey]. Similarly, an /u/ in a following syllable would cause labialization, particularly /a/ > [o], but occasionally also N > [y] and /e/ > [0] caused by a following semivowel. There was also an a-umlaut, which was a lowering of high vowels preceding an unstressed /a/. During the period from c. ad 500 to 700, called the 'syncopation period', Scandinavian underwent some important phonological changes, such as the loss of vowels in unstressed syllables. This loss led to the phonologization of certain allophonic variants. For example, the plural of land was phonologically *landu, pronounced with a labialized ('rounded') root vowel, *[londu]. When the final vowel was lost, the labialized root vowel became the mark of the plural for this class of nouns, and the [o] became a phoneme, written <g>.

In general, there are more umlaut effects in the West than in the East. In the eastern dialects of East Scandinavian there is no a-umlaut, and only a few traces of «-umlaut. The i-umlaut, however, seems to have extended throughout Scandinavia. All of these umlaut rules were productive at a period prior to that covered by our written records; therefore it is not possible to describe the rules accurately. The i-umlaut has great consequences for the inflectional morphology of the Scandinavian languages, and is the basis of important morphophonemic alterations, which will be treated in the section on morphology (pp. 45-53). It was - at least during a certain period - sensitive to syllable structure, therefore it did not apply in words with a short root syllable where the N was lost: stadr 'place* (< *stadiz). The a-umlaut has mainly affected the lexicon, and plays a less important role in the grammar of the languages.

One umlaut rule is still a synchronic rule of Old Norse, however, namely the so-called younger u-umlaut, which changes /a/ to /q/ in front of an unstressed /u/ in an inflectional ending, as in dggum, the dative plural of dagr 'day'. This rule is most consistently applied in Icelandic and in western Norwegian, less so in eastern Norwegian, and in East Scandinavian only in specific environments, such as across a nasal consonant.

Breaking is another effect of unstressed vowels on stressed root vowels. A-breaking would change a short /e/ in a root syllable to /ia/ under the influence of a following /a/, as in hjarta 'heart'. t/-breaking is the w-umlauted variant of this, caused by an original /u/ in the following syllable: jgrd 'earth' (< *erpu). By this process, initial /j/ was reintroduced into the language, after the loss of word-initial /j/ in Ancient Scandinavian.

Diphthongs

There are three diphthongs in Old Norse: /aei/, /qu/, /aey/. The first one has developed from Ancient Scandinavian /ai/ through a raising of the first element under the influence from the second (some kind of i-umlaut): /qu/ comes from /au/ through labialization of the first element under influence from the /u/ (some kind of m-umlaut); /aey/ is the /-umlaut of /au/. /«/ in /aey/ was furthermore labialized, and the diphthong developed into /0y/. In East Scandinavian the diphthongs were monophthongized early on: /ai/ > /aei/ > /e:/, /au/ > /qu/ > /0:/, /ey/ > /0y/ > /0:/. The trend started in Jutland and spread gradually east through Denmark and then north through southern and central Sweden and to parts of eastern Norway. By 1100 the diphthongs were monophthongized in all of Denmark and most of Sweden.

Vowels in Unstressed Syllables

The inventory of vowels in unstressed syllables is much smaller than that in stressed syllables. Instead of the sixteen phonemes of Table 3.1, there is only a contrast of three vowel phonemes in Old Norse, see Table 3.2. There is no length opposition, /a/ is distinguished from the other two by the feature [+ low]. The relevant feature is [± low] rather than [± high], which is shown by the fact that in many manuscripts, especially early Icelandic ones, the unstressed vowels are spelt <e> and <o> instead of <i> and <u>. /u/ is distinguished from the other two by the feature [+ labial]. This is shown by the fact that an unstressed /a/ becomes /u/ under w-umlaut, as in kglludu

Table 3.2 Vowels in unstressed syllables i u a

'called (3 pi.)', from kalla + du. (If the distinctive feature were [± back], there would be nothing for the w-umlaut to change.)

In Old Swedish and in eastern and northwestern dialects of Old Norwegian the use of -i/-w vs. -e/-o in unstressed syllables is determined by a principle of vowel harmony. Root syllables with a [+ high] vowel are followed by i and u in an unstressed syllable, as in flutti 'moved' and bitu 'bit (3 pi.)'; and root syllables with a [-high, -low] vowel are followed by e and o: dœmde 'judged, sentenced' and toko 'took (3 pi.)'. (After low root vowels the picture is less consistent.)

Consonants

The consonant phonemes of Old Norse are represented as in Table 3.3, where the consonants are given in the standard orthography. The non-strident non-sonorants form three groups of three consonants each: the labials, the velars, and the dentals [-labial, -velar]. Each of these has a voiceless stop (/p, t, k/), a voiced stop (/b, d, g/), and a fricative (/f, J), h/). In the labials and dentals the feature [± continuant] takes precedence over [± voice]; there is a voice opposition in the stops, and no voice opposition in the fricatives. That means that [f] and [v], and [6] and [ô] are in complementary distribution. The voiceless fricatives are used word initially, and the voiced ones word medially and finally. The letter <f> is used for both the voiced and the voiceless variant, as in fara [fara] 'go' and hafa [hava] 'have', whereas there are separate letters for the two dental allophones, as in padan 'thence'. In the velar series [+ voice] takes precedence over [± continuant]; there is a continuant opposition

Table 3.3 Consonant phonemes of Old Norse

P

b

f t

d

1>

k

9

h

s

m

n

r

1

Sonorant

_

_

_

_

_

_

_

+

+

+

+

Continuant

_

+

_

+

_

0

+

+

_

_

+

+

Velar

_

_

_ _

_

_

+

+

+

Labial

+

+

+

_

_

_

_

_

_

+

_

_

_

Strident

_

_

_ _

_

_

_

_

_

+

_

_

+

_

Voiced

-

+

0 -

+

0

-

+

-

-

+

+

+

+

between the two voiceless consonants, Ik/ and /h/, while the voiced /g/ may be a stop or a fricative depending on the environment.

By the Old Scandinavian period, the /h/ had been lost in all positions except word initially. Thus an /h/ which was the result of final devoicing of a fricative /g/ would also be lost: *mag > *mah > ma 'may, can*. In Norwegian and East Scandinavian the /h/ was lost everywhere except word initially before vowels and semivowels. This created a difference between Old Icelandic and Old Norwegian, as in hlutr 'part', hringr 'ring', hniga 'sink', etc. vs lutr, ringr, niga, etc.

A nasal preceding a final stop (which was devoiced) was generally assimilated to that stop in the West, but not in the East. Thus the past tense of binda 'bind' was bant (< band through final devoicing) in the East, but batt in the West.

A synchronic rule of Old Scandinavian is the assimilation of Ixl to a preceding /s/, /n/, or /l/. This takes place whenever a suffix starting with Ixl is added to a stem which ends in one of those consonants. In the case of l\l and In/ the rule does not apply after short stressed vowels: cf. stdll (< stol+r) 'table', ketill (< ketil+r) 'kettle', vs telr 'tells'. In most cases Ixml changes to Idl before /r/, as in madr (< mann+r) 'man'.

In Danish post-vocalic voiceless stops began to be voiced in the twelfth century, and later the voiced stops would develop into fricatives. Together with the vowel reduction mentioned above, this would lead to the characteristic Danish development: mata > made > made 'feed'. (The present-day orthography represents the middle stage.)

Towards the end of the thirteenth century certain consonant clusters began to be unacceptable, in particular final clusters ending in Ixl. Since the Ixl in most cases was an inflectional ending, it was not easily dropped. Instead, an epenthetic vowel was inserted, boendr > boender 'farmers'. The vowel would often be written <ae>. In western Norwegian and Icelandic an <u> was used.

Prosody

Stress

There is a distinction between stressed and unstressed syllables in Old Scandinavian. As we have already seen, the two kinds of syllables have a different inventory of distinctive vowel qualities. The stress is normally on the root syllable of a word; in most cases that is the first syllable. In compound words, the first element (or sometimes the second) has the primary stress, while the other element has a secondary stress. Certain prefixes may also have primary stress, in which case the root has secondary stress.

Quantity

Syllable quantity plays no significant part in the synchronic phonology of Old Scandinavian, but it did play a certain role in the derivational morphology of

Ancient Scandinavian (see below, pp.48 and 51), and it has far-reaching consequences for the subsequent development of the sound systems of the Scandinavian dialects. As we have seen, vowels may be short or long, and consonants may occur in clusters or be geminated. In stressed syllables, a short or a long vowel may be followed by none, one, or two (or more) consonants. Stressed syllables may thus be short, long, or 'overlong'.

In eastern Norwegian and western Swedish bisyllabic words have undergone certain phonological processes that are sensitive to the quantity of the root syllable, often referred to as vowel balance. On the one hand these processes have created new morphological patterns and distinctions in those dialects, and on the other hand they have set them off from the other Old Scandinavian dialects.

One such process is vowel reduction, which in these dialects affects only words with a long root syllable. After a long stressed syllable an unstressed vowel is reduced, while it is maintained after a short syllable. In eastern Norwegian this has led to the so-called 'cleft infinitive', with the ending -a after originally short root syllables (vera 'be') and -e after long root syllables (kaste 'throw'). In some of the Norwegian dialects the reduced vowel was completely dropped. In some words with a short root syllable the root vowel assimilated to the final vowel: gatu > gutu 'road'. The basis for these processes is the fact that a final syllable following a short root syllable receives some of the word stress, and is therefore better preserved. In some dialects such words probably had a 'balanced' stress.

In the further development of Scandinavian an important restructuring of the syllable structures took place. In West Scandinavian and Swedish an inter-dependency between stress and quantity arose; a stressed syllable had to be long. This means that the short stressed syllables were lengthened, either through lengthening of the vowel or through gemination of the consonant, depending on the actual consonants involved, and on the dialect. This change can be described as follows: The syllable boundary shifted towards the left, so that the last one of post-vocalic consonants can no longer count as part of the preceding syllable, and a long syllable is defined as a bi-moraic syllable. Thus a word like fe 'cattle' has still two morae, but now it counts as a long syllable and can still constitute a stressed syllable, /hoi/ 'hole' is reanalysed as /ho-1/ and becomes mono-moraic, therefore it changes into /ho:l/ or /holl/, and /koma/ 'come' might become /koima/ or /komma/. At the same time overlong syllables were also abolished, mostly through shortening of the vowel: natt > natt 'night'.

In Danish a different development took place; short vowels in stressed, open syllables were lengthened, /fara/ > /faire/ 'go, travel'. This did away with one type of short stressed syllables. On the other hand, all geminate consonants were shortened, pakk > tak 'thanks', which gave rise to a new type of short stressed syllables in monosyllabic words. In monosyllabic words with a short vowel plus a short consonant, the vowel would either remain short or be lengthened, as in /skip/ > /skiib/ 'ship'.

Tone

In most Norwegian and Swedish dialects there is today a distinction of two word tones in words of more than one syllable. These tones have never been recorded in writing, therefore we have only indirect evidence of their origin. The tonal difference was originally a difference between the pitch contour of monosyllabic and bisyllabic words. The two tones are therefore called 'single' (') and 'double' (") tone, respectively. In the modern languages there are also bisyllabic words with the single tone. These are mainly of three origins: they are loan words; they are monosyllabic roots with the definite article attached to them, /'baide/ (bad + et) 'the bath'; or they are words that have become bisyllabic through the insertion of an epenthetic vowel, /Inter/ (< bitr) 'bites (pres.)'. Words which were also bisyllabic in early Old Scandinavian have the double tone: /"bade/ 'bathe (inf.)', /"biter/ (< bitar) 'bites, bits (m. pi.)'. These facts indicate that the tonal distinction must have arisen before the definite article changed from being a clitic to becoming a suffix, and before the epenthetic vowel was introduced in final consonant clusters ending in an r, which means no later than early thirteenth century.

0 0

Post a comment